CAMShift Object Tracking Algorithm with Adaptive Window Sizing

This documentation presents the CAMShift (Continuously Adaptive Mean Shift) object tracking algorithm, which implements real-time target tracking with adaptive window sizing capability. The algorithm dynamically adjusts the tracking window dimensions based on changes in target size, ensuring optimized tracking performance under varying scale conditions. This implementation was personally developed for experimental purposes and demonstrates reliable performance with no identified bugs. The core functionality utilizes color probability distributions and mean shift iterations to locate target centroids, while incorporating window scaling mechanisms that respond to zeroth-order moments of the probability distribution. This algorithm can be effectively applied in computer vision systems for enhanced precision in object tracking and surveillance applications, featuring key functions including histogram back-projection, mean shift vector calculation, and moment-based window adaptation.